臺灣博碩士論文加值系統

近年來深度學習的議題迅速竄紅，深度學習解決了許多以前解決不了的問題，從熱門的議題，Google開發的AlphaGo、Facebook人臉辨識和Apple Siri語音辨識等，都是利用深度學習當主軸發展出來的新興產品，漸漸的深度學習變成一個在解決問題中不可或缺的角色。當中，影像處理的部分有很大的突破讓我們值得來加以探討。
本研究探討養寵物的比例逐漸增加，寵物的陪伴有效減輕心理壓力，甚至把寵物當作是家人般細心照顧與愛護，藉由寵物帶來的需求，各行各業紛紛創造寵物的新市場。因此，本論文提出一個基於深度學習之即時目標物尋找系統，使用目前在影像偵測中表現優異的深度學習架構進行偵測，以YOLOv2神經網路架構為基礎並加以設計訓練出最優權重，應用本系統之機器人，當使用者進行操作時能夠即時有效的偵測出寵物所在位置。

In recent years, deep learning has quickly become popularly and solving many before could not solve questions. From popular subject, Google developed Alpha Go, Facebook face recognition and Apple Siri voice recognition etc. Using deep learning to develop emerging products and gradually become an indispensable to solving problems. In the image processing has a breakthrough that makes us deserves to exploring.
The research explore the increasing proportion of keep a pet. Companionship with pet can Effective relieve psychological stress and taking care of like family. All working have creating new markets with the demand brought by pet. Therefore, the paper propose on real-time target searching system based on deep learning. Using the image processing of deep learning. Based on YOLOv2 network to training the new weights. Users can be operating immediately and effectively to detecting pet of location with system robot.

摘要 i
ABSTRACT ii
誌 謝 iii
目 錄 iv
表 目 錄 vi
圖 目 錄 vii
第 一 章 導 論 1
1.1. 研究動機 1
1.2. 研究目的 2
1.3. 論文架構 3
第 二 章 文 獻 探 討 4
2.1. 機器學習 4
2.1.1. 監督式學習 4
2.1.2. 非監督式學習 5
2.1.3. 半監督式學習 5
2.2. 類神經網路 (Neural Network, NN) 5
2.3. 深度學習 7
2.3.1. 卷積神經網路 (Convolutional Neural Network, CNN) 7
2.3.2. 深度置信網路 (Deep Belief Network, DBN) 11
2.3.3. 限制玻爾茲曼機 (Restricted Boltzmann Machine, RBM) 12
2.4. K-means演算法 13
2.5. 基礎物件偵測系統 14
2.5.1. Region-Based CNN (R-CNN) 14
2.5.2. You Only Look Once網路框架 15
2.5.3. YOLOv2網路框架 17
2.6. 深度學習框架 18
2.6.1. TensorFlow 18
2.6.2. Torch 18
2.6.3. Caffe 19
2.6.4. Cognitive Toolkit 19
第 三 章 研 究 方 法 20
3.1. 問題定義 20
3.2. 系統流程 21
3.3. 資料前置處理 22
3.4. YOLOv2資料訓練 27
3.4.1. YOLOv2設計 27
3.4.2. YOLOv2訓練新權重 28
3.5. 機器人建置地圖 31
3.6. 即時偵測系統 38
第 四 章 實 驗 結 果 與 討 論 41
4.1 實驗規劃 41
4.1.1 實驗環境 41
4.1.2 實驗資料 41
4.2目標物偵測率 42
4.3 移動即時偵測率 45
第 五 章 結 論 47
參 考 文 獻 48

[1] AlphaGo 李世乭五番棋 Web at
https://zh.wikipedia.org/wiki/AlphaGo%E6%9D%8E%E4%B8%96%E4%B9%AD%E4%BA%94%E7%95%AA%E6%A3%8B
[2] Goodfellow, Ian, et al. "Generative adversarial nets." Advances in neural
information processing systems.
[3] 機器學習的衰頹興盛：從類神經網路到淺層學習 Web at https://www.stockfeel.com.tw/%E6%A9%9F%E5%99%A8%E5%AD%B8%E7%BF%92%E7%9A%84%E8%A1%B0%E9%A0%B9%E8%88%88%E7%9B%9B%EF%BC%9A%E5%BE%9E%E9%A1%9E%E7%A5%9E%E7%B6%93%E7%B6%B2%E8%B7%AF%E5%88%B0%E6%B7%BA%E5%B1%A4%E5%AD%B8%E7%BF%92/
[4] Reducing the Dimensionality of Data with Neural Networks Web at
http://www.cs.toronto.edu/~hinton/science.pdf
[5] 類神經網路的復興：深度學習簡史Web at
https://www.stockfeel.com.tw/%E9%A1%9E%E7%A5%9E%E7%B6%93%E7%B6%B2%E8%B7%AF%E7%9A%84%E5%BE%A9%E8%88%88%EF%BC%9A%E6%B7%B1%E5%BA%A6%E5%AD%B8%E7%BF%92%E7%B0%A1%E5%8F%B2/
[6] 一文讀懂卷積神經網絡CNN Web at
https://read01.com/zh-tw/7Rx00O.html#.W06ubdIzaUl
[7] 卷積神經網絡及其在圖像處理中的應用 Web at https://read01.com/oyPn3.html
[8] Fukushima, Kunihiko. "Neural Network Model for a Mechanism of Pattern
Recognition Unaffected by Shift in Position- Neocognitron." Electron. & Commun. Japan 62.10 (1979): 11-18.
[9] Boureau, Y-Lan, Jean Ponce, and Yann LeCun. "A theoretical analysis of feature
pooling in visual recognition." Proceedings of the 27th international conference on machine learning (ICML-10). 2010.
[10] Shaoqing Ren, Kaiming He, Ross Girshick, Jian Sun, 2015, “Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks”.
[11] 關於影像辨識，所有你應該知道的深度學習模型 Web at
https://medium.com/@syshen/%E7%89%A9%E9%AB%94%E5%81%B5%E6%B8%AC-object-detection-740096ec4540
[12] Joseph Redmon, Santosh Divvala, Ross Girshick, Ali Farhadi, 2015,” You Only
Look Once: Unified, Real-Time Object Detection”.
[13] Joseph Redmon, Ali Farhadi,2016,” YOLO9000:Better, Faster, Stronger”.
[14] Abadi, Martín, et al. "Tensorflow: Large-scale machine learning on
heterogeneous distributed systems." arXiv preprint arXiv:1603.04467 (2016).
[15] 機器學習開源框架 Web at https://read01.com/nJnKnJ.html
[16] 深度學習—你需要了解的八大開源框架 Web at https://read01.com/zh-tw/Ka3R5d.html#.W0djZtIza5t
[17] Yu, Dong, et al. "An introduction to computational networks and the computational network toolkit." Microsoft Technical Report MSR-TR-2014–112 (2014).
[18] 機器學習: 集群分析 K-means Clustering Web at
https://medium.com/@chih.sheng.huang821/%E6%A9%9F%E5%99%A8%E5%AD%B8%E7%BF%92%E9%9B%86%E7%BE%A4%E5%88%86%E6%9E%90-k-means-clustering-e608a7fe1b43